You can do every statistical 'right thing' and still be wrong if you ask a bad question or ignore history and causality. Good analysis needs aesthetic judgment—taste about questions, variables, and narratives—beyond tidy charts, p‑values, and reviewer‑pleasing formatting. Packaging can hide artless thinking that should be rejected.
— This challenges rule‑based peer review and training by arguing institutions must reward causal judgment and domain knowledge, not just methodological hygiene.
Sebastian Jensen
2025.08.30
72% relevant
The post contends preregistration, open code, and multiple-testing fixes don't reliably stop bad research and instead burden good work, echoing the idea that checklists can’t substitute for judgment and causal taste in analysis.
José Duarte
2025.08.28
78% relevant
Duarte’s critique of Napier & Jost (relabeling attitude items as 'rationalization') and Lewandowsky et al. (linking moon‑hoax beliefs with climate denial despite 10/1145 endorsers) shows that methodological 'hygiene' can still yield nonsense if constructs and base rates make the inference incoherent; judgment and logic must precede measurement.
Paul S
2025.08.28
78% relevant
By showing that technically rigorous theorizing under Rawls’s 'reasonable agents' and 'strict compliance' assumptions can be irrelevant to actual politics, the piece echoes the claim that method without good question selection and domain judgment yields impressive but useless work.
Sebastian Jensen
2025.08.27
100% relevant
His 22‑country regression linking 2005 GDP to Jewish population share (with code and proper diagnostics) nevertheless yielded a wrong 'Jews drive socialism' story he now disowns as historically ignorant.
Santa Fe Institute
2024.10.28
45% relevant
Krakauer’s beauty–interesting frame parallels the claim that good analysis requires aesthetic judgment beyond procedural checklists; elegant ('beautiful') models must be balanced with attention to 'interesting' organized complexity in real systems.